Perbandingan Kinerja Algoritma Klasifikasi untuk Pemilihan Tempat Promosi upaya Meningkatkan Jumlah Mahasiswa Baru

Abstract

Bahasa Indonesia

Efektivitas strategi promosi menjadi kunci keberhasilan institusi pendidikan tinggi dalam menarik minat calon mahasiswa baru. Namun, banyak institusi menghadapi masalah serius dalam pemilihan lokasi promosi yang tepat, yang mengakibatkan penurunan minat calon mahasiswa dan pemborosan sumber daya. Untuk mengatasi masalah ini, penelitian ini mengembangkan model klasifikasi berbasis data mining yang mampu mengidentifikasi lokasi promosi paling efektif. Tiga algoritma klasifikasi yang digunakan dalam penelitian ini adalah Logistic Regression, Support Vector Machine (SVM), dan Decision Tree (C4.5). Data pendaftaran mahasiswa dikumpulkan dan diproses melalui tahapan pra pemprosesan yang meliputi penggantian nilai hilang, normalisasi data, dan transformasi atribut nominal menjadi numerik. Data kemudian dibagi menjadi subset pelatihan dan pengujian menggunakan metode split data dengan rasio 70:30. Hasil evaluasi menunjukkan bahwa model Decision Tree (C4.5) memberikan performa terbaik dengan accuracy 93.75%, precision 97.37%, dan recall 90.24%. Logistic Regression juga menunjukkan hasil yang memuaskan dengan accuracy 90.00%, precision 92.31%, dan recall 87.80%. Sementara itu, SVM menunjukkan performa yang lebih rendah dengan accuracy 72.50%, precision 80.65%, dan recall 60.98%. Kesimpulannya, model Decision Tree (C4.5) dan Logistic Regression dapat diandalkan untuk mengoptimalkan strategi promosi institusi pendidikan tinggi, memastikan alokasi sumber daya yang lebih efisien dan efektif, serta meningkatkan jumlah pendaftar baru. Penelitian ini juga memberikan kontribusi signifikan dalam literatur terkait penggunaan data mining untuk strategi promosi di sektor pendidikan tinggi. Abstract The effectiveness of promotional strategies is crucial for higher education institutions in attracting new student enrollments. However, many institutions face serious issues in selecting the appropriate promotional locations, leading to decreased student interest and resource wastage. To address this issue, this study develops a data mining-based prediction model capable of identifying the most effective promotional locations. The three classification algorithms used in this study are Logistic Regression, Support Vector Machine (SVM), and Decision Tree (C4.5). Student enrollment data were collected and processed through pre-processing stages, including missing value replacement, data normalization, and transformation of nominal attributes to numerical. The data were then split into training and testing subsets using a 70:30 split ratio. Evaluation results indicate that the Decision Tree (C4.5) model performed the best with an accuracy of 93.75%, precision of 97.37%, and recall of 90.24%. Logistic Regression also showed satisfactory results with an accuracy of 90.00%, precision of 92.31%, and recall of 87.80%. Meanwhile, SVM demonstrated lower performance with an accuracy of 72.50%, precision of 80.65%, and recall of 60.98%. In conclusion, the Decision Tree (C4.5) and Logistic Regression models are reliable for optimizing promotional strategies of higher education institutions, ensuring more efficient and effective resource allocation, and increasing new student enrollments. This study also makes a significant contribution to the literature related to the use of data mining for promotional strategies in the higher education sector.

English

Efektivitas strategi promosi menjadi kunci keberhasilan institusi pendidikan tinggi dalam menarik minat calon mahasiswa baru. Namun, banyak institusi menghadapi masalah serius dalam pemilihan lokasi promosi yang tepat, yang mengakibatkan penurunan minat calon mahasiswa dan pemborosan sumber daya. Untuk mengatasi masalah ini, penelitian ini mengembangkan model klasifikasi berbasis data mining yang mampu mengidentifikasi lokasi promosi paling efektif. Tiga algoritma klasifikasi yang digunakan dalam penelitian ini adalah Logistic Regression, Support Vector Machine (SVM), dan Decision Tree (C4.5). Data pendaftaran mahasiswa dikumpulkan dan diproses melalui tahapan pra pemprosesan yang meliputi penggantian nilai hilang, normalisasi data, dan transformasi atribut nominal menjadi numerik. Data kemudian dibagi menjadi subset pelatihan dan pengujian menggunakan metode split data dengan rasio 70:30. Hasil evaluasi menunjukkan bahwa model Decision Tree (C4.5) memberikan performa terbaik dengan accuracy 93.75%, precision 97.37%, dan recall 90.24%. Logistic Regression juga menunjukkan hasil yang memuaskan dengan accuracy 90.00%, precision 92.31%, dan recall 87.80%. Sementara itu, SVM menunjukkan performa yang lebih rendah dengan accuracy 72.50%, precision 80.65%, dan recall 60.98%. Kesimpulannya, model Decision Tree (C4.5) dan Logistic Regression dapat diandalkan untuk mengoptimalkan strategi promosi institusi pendidikan tinggi, memastikan alokasi sumber daya yang lebih efisien dan efektif, serta meningkatkan jumlah pendaftar baru. Penelitian ini juga memberikan kontribusi signifikan dalam literatur terkait penggunaan data mining untuk strategi promosi di sektor pendidikan tinggi. Abstract The effectiveness of promotional strategies is crucial for higher education institutions in attracting new student enrollments. However, many institutions face serious issues in selecting the appropriate promotional locations, leading to decreased student interest and resource wastage. To address this issue, this study develops a data mining-based prediction model capable of identifying the most effective promotional locations. The three classification algorithms used in this study are Logistic Regression, Support Vector Machine (SVM), and Decision Tree (C4.5). Student enrollment data were collected and processed through pre-processing stages, including missing value replacement, data normalization, and transformation of nominal attributes to numerical. The data were then split into training and testing subsets using a 70:30 split ratio. Evaluation results indicate that the Decision Tree (C4.5) model performed the best with an accuracy of 93.75%, precision of 97.37%, and recall of 90.24%. Logistic Regression also showed satisfactory results with an accuracy of 90.00%, precision of 92.31%, and recall of 87.80%. Meanwhile, SVM demonstrated lower performance with an accuracy of 72.50%, precision of 80.65%, and recall of 60.98%. In conclusion, the Decision Tree (C4.5) and Logistic Regression models are reliable for optimizing promotional strategies of higher education institutions, ensuring more efficient and effective resource allocation, and increasing new student enrollments. This study also makes a significant contribution to the literature related to the use of data mining for promotional strategies in the higher education sector.

Keywords

• Data Mining
• Klasifikasi
• Decision Tree
• Logistic Regression
• Promosi Pendidikan